Leveraging Open Data to Support Forest Mapping, Modeling, and Policy Analysis in the Pacific Northwest, USA

Abstract

Increasing access to open data can offer unprecedented opportunities to conduct and share actionable science involving forest ecology, management, and conservation. This dissertation recounts three diverse applications of open data and ecosystem data science in the Pacific Northwest, USA: (i) to generate maps of forest conditions that can be easily updated as new earth observation data arrive; (ii) to develop a reproducible workflow to recalibrate the Forest Vegetation Simulator (FVS), a widely used growth-and-yield model, as new field observations arise; and (iii) to quantify tradeoffs in timber, carbon, and cash flow among alternative management approaches in Douglas-fir forests. Compiling a multi-agency dataset across Oregon and Washington, I demonstrate the potential for modern Machine Learning algorithms to effectively learn and generalize relationships between forest structure attributes with remote sensing across multiple ecoregions. Utilizing an extensive network of repeated forest inventory measurements, I highlight opportunities to simplify FVS equations and software maintenance while improving predictions of diameter growth for sixteen diverse tree species and learning growth drivers simultaneously across multiple ecoregions from coastal Washington through California. Finally, by translating an open dataset depicting forest conditions across the landscape to stand-level units, I employ more than 250,000 simulations of forest management across western Oregon and Washington to conduct a tradeoff analysis involving varying rotation lengths, levels of green tree retention, and extent of riparian buffers. I find that forest management following Forest Stewardship Council requirements is likely to produce substantially higher carbon storage in forests and in wood products at a significant opportunity cost in terms of net present value of timber revenue compared to business-as-usual management following state forest practice regulations in both Oregon and Washington. Together, these three applications illustrate common use cases for forest data and technology to enhance decision-making capacity and understanding of current forest conditions, how confidently we can model forest dynamics, and how alternative management approaches can lead to important tradeoffs relevant to forest owners, managers, policymakers, and consumers of forest products.